20,371 research outputs found
Metadata Representations for Queryable ML Model Zoos
No full textMachine learning (ML) practitioners and organizations are building model zoos of pre-trained models, containing metadata describing properties of the ML models and datasets that are useful for reporting, auditing, reproducibility, and interpretability purposes. The metatada is currently not standardised; its expressivity is limited; and there is no interoperable way to store and query it. Consequently, model search, reuse, comparison, and composition are hindered. In this paper, we advocate for standardized ML model metadata representation and management, proposing a toolkit supported to help practitioners manage and query that metadata.Web Information SystemsHuman-Centred Artificial Intelligenc
A Manifesto of Nodalism
No full textThis paper proposes the notion of Nodalism as a means describing contemporary culture and of understanding my own creative practice in electronic music composition. It draws on theories and ideas from Kirby, Bauman, Bourriaud, Deleuze, Guatarri, and Gochenour, to demonstrate how networks of ideas or connectionist neural models of cognitive behaviour can be used to contextualize, understand and become a creative tool for the creation of contemporary electronic music
Optimizing ML Inference Queries Under Constraints
No full textThe proliferation of pre-trained ML models in public Web-based model zoos facilitates the engineering of ML pipelines to address complex inference queries over datasets and streams of unstructured content. Constructing optimal plan for a query is hard, especially when constraints (e.g. accuracy or execution time) must be taken into consideration, and the complexity of the inference query increases. To address this issue, we propose a method for optimizing ML inference queries that selects the most suitable ML models to use, as well as the order in which those models are executed. We formally define the constraint-based ML inference query optimization problem, formulate it as a Mixed Integer Programming (MIP) problem, and develop an optimizer that maximizes accuracy given constraints. This optimizer is capable of navigating a large search space to identify optimal query plans on various model zoos.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Web Information SystemsHuman-Centred Artificial Intelligenc
Anemia screening via non-invasive transcutaneous measurement of hematocrit in hemodialysis and pre-ESRD patients
No full textCELLULAR BASIS OF DIABETIC NEPHROPATHY: IV ANTIOXIDANT ENZYME MRNA EXPRESSION LEVELS IN SKIN FIBROBLASTS OF TYPE 1 DIABETIC SIBLING PAIRS.
No full textBACKGROUND:
Blunted cultured skin fibroblast (SF) antioxidant enzyme responses to hyperglycaemia are associated with diabetic nephropathy risk. The present study explores whether this association is, at least in part, genetically determined.
METHODS:
We measured glomerular structure and SF mRNA expression for catalase and glutathione peroxidase in 21 sibling pairs concordant for type 1 diabetes. All patients had four or more (mean 21.5) years of diabetes and glomerular filtration rate>40 ml/min/1.73 m2. Thirty-four patients were normoalbuminuric, four were microalbuminuric, three were proteinuric and one was not classifiable. Heritability of patient characteristics was assessed by intra-class correlation and by a genetic variance component model.
RESULTS:
Mesangial fractional volume, mesangial matrix fractional volume, glomerular basement membrane width and surface density of peripheral glomerular basement membrane per glomerulus were significantly correlated in these sibling pairs. Catalase mRNA expression levels were also related and highly heritable in these sibling pairs. The association between sibship and glutathione peroxidase mRNA expression levels did not reach statistical significance.
CONCLUSIONS:
This study suggests that SF catalase mRNA expression levels, known to be associated with diabetic nephropathy risk, are in part genetically determined
Lophotermes aduncus Miller
No full textLophotermes aduncus Miller (Figs 17 –20, 65) Lophotermes aduncus Miller, 1991: 1212. Type material examined: paratypes. AUSTRALIA. ANIC. Queensland. # 10 –20450, 8 km SW of Cape Weymouth, 1.vii. 1984, L.R. Miller. Other material examined. PAPUA NEW GUINEA. Authors' collection. Fly. # PNGT 1412: Morehead, 22.iii. 1989, YR & ML. # PNGT 1455: Morehead, 25.iii. 1989, YR & ML. # PNGT 1469: Morehead, 27.iii. 1989, YR & ML. Diagnosis. L. aduncus can be differentiated from congeners by the narrow frontal projection and lateral tubercles, and the short mandibles of its soldiers. Worker enteric valve with 6 cushions of equal size, with many small spines and long, curved ones in their distal part. For a detailed description see Miller (1991). Measurements as in Table 7 and 18. Distribution (Fig. 65). L. aduncus was collected three times at Morehead, in Southern Papua New Guinea savannas. It is also known from various places in Queensland (Miller 1991, Watson & Abbey 1993: 110). Head length to apex of frontal tubercle 1.52–1.91 Head length to anterolateral corner of the genae 1.47–1.66 Head depth with postmentum 0.76–0.91 Head depth without postmentum 0.86–1.01 Head maximum width 1.10–1.19 Pronotum width 0.59–0.66 Left mandible length 1.54–1.74 Postmentum minimum width 0.14–0.20 Distance between tips of lateral tubercles 0.70–0.81Published as part of Bourguignon, Thomas, Leponce, Maurice & Roisin, Yves, 2008, Revision of the Termitinae with snapping soldiers (Isoptera: Termitidae) from New Guinea, pp. 1-34 in Zootaxa 1769 on page 12, DOI: 10.5281/zenodo.18209
Frontal white matter tracts sustaining speech production in primary progressive aphasia
No full textIn primary progressive aphasia (PPA), speech and language difficulties are caused by neurodegeneration of specific brain networks. In the nonfluent/agrammatic variant (nfvPPA), motor speech and grammatical deficits are associated with atrophy in a left fronto-insular-striatal network previously implicated in speech production. In vivo dissection of the crossing white matter (WM) tracts within this "speech production network" is complex and has rarely been performed in health or in PPA. We hypothesized that damage to these tracts would be specific to nfvPPA and would correlate with differential aspects of the patients' fluency abilities. We prospectively studied 25 PPA and 21 healthy individuals who underwent extensive cognitive testing and 3 T MRI. Using residual bootstrap Q-ball probabilistic tractography on high angular resolution diffusion-weighted imaging (HARDI), we reconstructed pathways connecting posterior inferior frontal, inferior premotor, insula, supplementary motor area (SMA) complex, striatum, and standard ventral and dorsal language pathways. We extracted tract-specific diffusion tensor imaging (DTI) metrics to assess changes across PPA variants and perform brain-behavioral correlations. Significant WM changes in the left intrafrontal and frontostriatal pathways were found in nfvPPA, but not in the semantic or logopenic variants. Correlations between tract-specific DTI metrics with cognitive scores confirmed the specific involvement of this anterior-dorsal network in fluency and suggested a preferential role of a posterior premotor-SMA pathway in motor speech. This study shows that left WM pathways connecting the speech production network are selectively damaged in nfvPPA and suggests that different tracts within this system are involved in subcomponents of fluency. These findings emphasize the emerging role of diffusion imaging in the differential diagnosis of neurodegenerative diseases
Building a generalisable ML pipeline at ING
No full textAdvances in data science have caused an increase in the use of Artificial Intelligence (AI), specifically Machine Learning (ML), throughout various fields. Not only in research but in the industry as well, has ML been receiving increasing amounts of interest. Many companies rely on ML models to increase the efficiency of existing processes or offer new services and products. The industry, however, is facing several additional challenges compared to the academic context. One of those challenges is applying the Development Operations (DevOps) model to an ML application, also referred to as MLOps. This thesis sets out to find the specific challenges that practitioners encounter while operationalising ML models. To do so, we perform a single-case case study on an ML pipeline built by the Trade & Communication Surveillance team at the ING bank. This case study consists of conducting a set of interviews and performing a manual code inspection of the pipeline. The team faces challenges ranging from having insufficient time for operationalising each ML project individually to operating in the highlyregulated fintech context. Their pipeline is able to deploy a single ML model but it does not generalise well to other projects. We present the first version of an application that mitigates these challenges. The application is able to deploy ML models to the development environment at ING and can be operated by data scientists to reduce the effort of operationalising an ML model. Computer Science | Software Technolog
Thermalized (350K) QM7b, GDB-13, water, and short alkane quantum chemistry dataset including MOB-ML features
No full textMOB-ML features, HF energies, pair correlation energies, and geometries. Data are provided for water, a dataset of alkanes, the QM7b dataset, and a subset of GDB13.
File updated 2019-09-04 to add missing energy.dat to qm7b_T folderRelated Publication:
A Universal Density Matrix Functional from Molecular Orbital-Based Machine Learning: Transferability across Organic Molecules
Cheng, Lixue Caltech
Welborn, Matthew Caltech
Christensen, Anders S. Universität Basel
Miller, Thomas F., III Caltech
arXiv:1901.03309
2019-01-10
en
Ephelotermes paleatus Miller
No full textEphelotermes paleatus Miller (Figs 9 –12, 47–48, 66) Ephelotermes paleatus Miller, 1991: 1188. Type material examined: paratypes from type locality. AUSTRALIA. Queensland. Weipa, 9.xi. 1978, L.R. Miller. ANIC # 10–18368. Other material examined. PAPUA NEW GUINEA. Authors' collection. Fly. # PNGT 1467: Morehead, 27.iii. 1989, YR & ML. # PNGT 1578: Lake Murray, 24.v. 1990, YR & ML. Diagnosis. This species is easily recognised from congeners by its small size (Tables 3 –4, 18) as well as by the downward projected postmentum of soldiers (Fig. 10). For additional description see Miller (1991). New Guinean specimens of this species slightly differ from Australian ones by the shape of the frontal projection, which is directed upwards. However, we judge this character insufficient to consider them as a distinct species. Distribution (Fig. 66). We collected this species twice in Papua New Guinea, at Morehead and Lake Murray (Fly). It is also known from northern Queensland, Australia (Miller 1991, Watson & Abbey 1993: 102). Length without wings 10.58 Head length to apex of the labrum 0.87 Head length to clypeo-frontal suture 0.50 Head width with eyes 0.74 Eye maximum diameter 0.20 Ocellus maximum length 0.10 Pronotum length 0.45 Pronotum width 0.70 Head length to apex of frontal tubercle 0.99–1.12 Head length to anterolateral corner of the genae 0.91–1.02 Head depth with postmentum 0.45–0.57 Head depth without postmentum 0.57–0.71 Head maximum width 0.60–0.72 Pronotum width 0.38–0.46 Left mandible length 1.24–1.33 Postmentum minimum width 0.14–0.21Published as part of Bourguignon, Thomas, Leponce, Maurice & Roisin, Yves, 2008, Revision of the Termitinae with snapping soldiers (Isoptera: Termitidae) from New Guinea, pp. 1-34 in Zootaxa 1769 on page 8, DOI: 10.5281/zenodo.18209
- …
